Using Airborne Measurements to Evaluate HRRR Forecasts of Freezing Drizzle Aloft: Results from the WINTRE-MIX Field Campaign

Supercooled large drop (SLD) icing poses a serious threat to aviation, however, forecasting SLD occurrence is challenging, due in part to limitations in numerical weather prediction models. One goal of the Winter Precipitation Type Research Multi-scale Experiment (WINTRE-MIX) was the evaluation and improvement of numerical forecasts of precipitation type (p-type). Research flights were conducted during WINTRE-MIX with the National Research Council of Canada Convair-580 aircraft to investigate precipitation processes in winter storms with near-freezing surface conditions. The High-Resolution Rapid Refresh model (HRRR) is commonly used in the meteorological community for its 3-km resolution along with its short temporal data assimilation of one hour, thus, making it an ideal model for mesoscale events such as winter storms. The HRRR uses the Thompson–Eidhammer bulk microphysics scheme to represent clouds and precipitation processes. This study investigates a WINTRE-MIX intensive observing period that took place on 07–08 March 2022 that coincided with a warm frontal passage through northern New York and into southern Quebec from 1600 to 2000 UTC. During initial ascent, the Convair observed widespread freezing drizzle (FZDZ) with cloud-top temperatures as cold as -15℃. Cloud-top temperatures observed with FZDZ are typically greater than -12℃, making this flight of interest from a microphysical perspective.

The aim of this study is to evaluate HRRR forecasts against Convair-580 observations. To characterize icing conditions, ice detector frequency, liquid water content (LWC), cloud droplet number concentration, hydrometeor particle size distribution, and maximum droplet diameter (D_max) are examined and compared between in situ aircraft observations to their model-simulated equivalent. Using the method presented by Tessendorf et al. (2021), D_max is calculated using HRRR data and compared to measurements from 2D-Stereo probe. The inclusion of D_max allows for the distinction between freezing rain (FZRA) and FZDZ. Additionally, we compare the observed and simulated occurrence of ice particles to evaluate HRRR’s ability to predict the transition between FZDZ and snow. Thermodynamic conditions and cloud structure are compared between aircraft and model simulations to evaluate the capabilities of the HRRR to represent the observed development of SLD. These comparisons are supported by the use of airborne W-band and X-band radar profiles collected by the Convair-580 along with soundings made across the WINTRE-MIX domain.